Supply chain transformational leadership and resilience: the mediating role of ambidextrous business model

Frequent supply chain disruptions during the COVID-19 pandemic have underscored the need for firms to build supply chain resilience (SCR), defined as the capability of a system to maintain desirable functions before/during disruptions and recover to normal after disruptive events. While prior work largely emphasizes resources and capabilities (e.g., agility, redundancy, collaboration), the strategic role of supply chain transformational leadership (SCTL)—the focal firm’s influence that models values and reformative behaviors to inspire and develop close relationships with supply chain partners—has been underexplored. Drawing on social learning theory, the authors posit that a focal firm exhibiting high SCTL can serve as a role model whose transformative behaviors are emulated by partners, thereby enhancing SCR. The study also argues that firms face exploration–exploitation tensions and proposes that an ambidextrous business model (balancing novelty and efficiency) is a mechanism through which SCTL improves SCR. Moreover, paradox cognition (the capacity to recognize and embrace contradictory demands) is proposed to strengthen SCTL’s effect on ambidextrous business model design. The study addresses three research questions: (1) Is SCTL positively related to SCR? (2) Does the ambidextrous business model mediate the SCTL–SCR relationship? (3) Does paradox cognition strengthen the effect of SCTL on the ambidextrous business model?

Supply chain resilience (SCR) has roots in engineering, ecology, and psychology and is variably defined as the ability to resist and/or rebound from disruptions. It can be conceptualized at the firm or supply chain system level and decomposed into dimensions; following Cheng and Lu (2017), this study adopts proactive SCR (capability to mitigate shocks and maintain functions before/during disruptions) and reactive SCR (capability to respond and recover after disruptions). Prior antecedents of SCR cluster into reengineering, collaboration, agility, and risk management culture, with some work on mixed effects including Industry 4.0, social capital, leadership, and business model innovation. Transformational leadership, comprising inspiration, intellectual stimulation, and individualized consideration, can, via social learning processes, enhance both proactive and reactive SCR by motivating, building trust, and coordinating behaviors across partners (H1a, H1b). Leadership is also linked to organizational ambidexterity. An ambidextrous business model balances novelty-centered design (new governance/content/structure to create value) with efficiency-centered design (reconfiguring existing elements to improve efficiency). SCTL can foster a supportive organizational context that encourages such ambidextrous design (H2). Ambidextrous business models are expected to improve both proactive and reactive SCR by guiding partners to maintain agility, efficiency, and rapid response (H3a, H3b). Thus, ambidextrous business model is hypothesized to mediate SCTL’s effects on both SCR dimensions (H4a, H4b). Paradox cognition should amplify SCTL’s influence on ambidextrous business model by helping firms recognize and embrace exploration–exploitation tensions (H5). A conceptual model links SCTL to SCR directly and indirectly via ambidextrous business model, with paradox cognition moderating the SCTL–ambidextrous business model path.

Research design and data collection: The study surveyed Chinese manufacturing firms, an appropriate context due to significant COVID-19 disruptions and China’s central role in global supply chains. To capture regional diversity, firms were sampled from Guangdong, Jiangsu, Shandong (eastern coastal, high economic force/transport), Henan (middle, average), and Inner Mongolia (northwest, lower levels). Questionnaire development followed three steps: executive interviews (n=12) to validate content; literature-based item development with translation/back-translation; and a pre-test with 20 executives. To mitigate common method bias (CMB), the survey was split into two parts completed by different respondents within each firm: Part A included demographics, competitive intensity, SCTL, novelty-centered business model, and SCR; Part B included paradox cognition and efficiency-centered business model. Data were collected via email from May to December 2020. Of 435 firms agreeing to participate, 317 valid firm-level responses were obtained (72.9% effective response): Guangdong (72), Jiangsu (62), Shandong (67), Henan (56), Inner Mongolia (60). Respondents averaged 7.19 years’ experience; 64.8% were CEOs/general managers/vice presidents; 35.2% were operations directors. Non-response bias checks (size, ownership comparisons; early vs. late responses across size, age, industry, ownership) showed no significant bias. Measures: All constructs used seven-point Likert scales (1=strongly disagree, 7=strongly agree). SCTL was measured with a refined 7-item scale (Defee et al., 2010). Paradox cognition used a 7-item scale (Smith & Lewis, 2011). Ambidextrous business model was operationalized via novelty-centered business model (10 items) and efficiency-centered business model (9 items) adapted from Zott & Amit (2007); their average formed the ambidexterity index. SCR was measured using two 4-item scales for proactive and reactive dimensions (Ambulkar et al., 2015; Brandon-Jones et al., 2014; Wieland & Wallenburg, 2013). Controls included firm size (ln employees), firm age (ln years since founding), industry (dummy: high-tech=1), ownership (state-owned/collective, private; foreign-invested as baseline), and competitive intensity (4 items from Jaworski & Kohli, 1993). Reliability and validity: Cronbach’s alpha values exceeded 0.7 for all constructs. Exploratory factor analysis extracted seven components consistent with constructs. Confirmatory factor analysis (AMOS 24.0) showed good model fit: χ²/df=2.034; RMSEA=0.057; CFI=0.928; NNFI=0.923; SRMR=0.038. Composite reliabilities exceeded 0.7; item loadings ranged 0.760–0.939; AVEs exceeded 0.5; discriminant validity was supported (correlations below square roots of AVEs). Common method bias checks included Harman’s single-factor test (no dominant factor), comparison of one-factor vs. seven-factor CFA (one-factor worse), and adding a common method factor (fit unchanged), indicating no serious CMB. Analysis: Hypotheses were tested using hierarchical regression (SPSS 23.0), with mean-centered interaction terms for moderation (Aiken & West, 1991). Multicollinearity was low (max VIF=1.739). Mediation was assessed via Baron and Kenny steps and bootstrapping (PROCESS macro; 5000 samples). Simple slope analysis probed moderation.

- SCTL positively affects SCR: Proactive dimension β=0.122, p<0.05; Reactive dimension β=0.166, p<0.01 (H1a, H1b supported). - SCTL positively affects ambidextrous business model: β=0.140, p<0.05 (H2 supported). - Ambidextrous business model positively affects SCR: Proactive β=0.241, p<0.001; Reactive β=0.256, p<0.001 (H3a, H3b supported). - Mediation by ambidextrous business model: • For Proactive SCR, full mediation: when ambidexterity is included, its effect is β=0.228, p<0.001 and SCTL becomes non-significant (β=0.090, p>0.1). Bootstrapped indirect effect=0.032 (Boot SE=0.017; 95% CI [0.004, 0.070]) (H4a supported). • For Reactive SCR, partial mediation: ambidexterity β=0.237, p<0.001; SCTL remains significant β=0.133, p<0.05. Bootstrapped indirect effect=0.033 (Boot SE=0.017; 95% CI [0.005, 0.069]) (H4b supported). - Moderation by paradox cognition: SCTL × Paradox cognition interaction on ambidextrous business model is positive and marginally significant β=0.094, p<0.1; simple slope analysis shows the SCTL→ambidextrous business model relationship is stronger at higher paradox cognition (H5 supported). - Robustness: PROCESS bootstrapping results are consistent with hierarchical regressions. Measurement model fit indices indicate reliable and valid constructs (χ²/df=2.034; RMSEA=0.057; CFI=0.928; NNFI=0.923; SRMR=0.038). Multicollinearity was not a concern (max VIF=1.739).

Findings confirm that SCTL enhances both proactive and reactive SCR through social learning mechanisms that promote shared vision, trust, coordination, and emulation of resilient behaviors across supply chain partners. The ambidextrous business model is a key mechanism: SCTL fosters a supportive context (discipline, stretch, trust) that enables balancing novelty-centered and efficiency-centered designs, which in turn improves supply chain agility, robustness, and responsiveness. The mediation pattern—full for proactive and partial for reactive—suggests that proactively mitigating shocks relies more heavily on ambidextrous business model design, whereas post-disruption recovery also benefits directly from SCTL’s inspirational and individualized support. Paradox cognition strengthens SCTL’s influence on ambidextrous business model by helping firms recognize and embrace exploration–exploitation tensions, thereby increasing the effectiveness of leadership-driven ambidexterity. Overall, the study advances a leadership-centered view of building SCR, complementing resource/capability perspectives by highlighting role-modeling and organizational learning processes across the supply chain.

Drawing on social learning theory, the study demonstrates that SCTL positively influences both proactive and reactive SCR. An ambidextrous business model fully mediates the SCTL→proactive SCR relationship and partially mediates the SCTL→reactive SCR relationship, while paradox cognition strengthens the SCTL→ambidextrous business model link. These results elucidate how leadership behaviors propagate through supply chain networks to enhance resilience via ambidextrous business model design. The study contributes theoretically by expanding antecedents of SCR to include SCTL, opening the ‘black box’ via ambidextrous business model as a mechanism, and identifying paradox cognition as a boundary condition. Practically, it suggests that managers should cultivate transformational leadership, design ambidextrous business models balancing novelty and efficiency, and foster paradox cognition to enhance SCR.

- Scope of leadership: The study focuses on SCTL; future work could examine other leadership styles (e.g., transactional leadership) to broaden antecedents of SCR. - Mechanisms: Only ambidextrous business model is tested as a mediator; future research should explore additional pathways and configurational approaches. - Boundary conditions: Paradox cognition is one moderator; further moderators (e.g., environmental dynamism) and moderated mediation models should be investigated to map contingencies more fully.